一种应用于半导体制造业的支持向量机SVM检测方法  

A Detection Method of Support Vector Machine SVM for Semiconductor Manufacturing

作　　者：王艳生 俞微 魏峥颖 WANG Yansheng;YU Wei;WEI Zhengying(Shanghai Huali Microelectronics Corporation,Shanghai 201203,China.)

机构地区：[1]上海华力微电子有限公司,上海201203

出　　处：《集成电路应用》2019年第8期56-59,共4页Application of IC

基　　金：上海市经济和信息化委员会软件和集成电路产业发展专项基金(1500204)

摘　　要：提出一种支持向量机SVM的检测方法,用于半导体制造过程中特征参数检测。首先对SVM模型进行分析,结果显示2D SVM和模糊SVM模型效果最佳。然后提出一种将2D SVM和模糊SVM这两种模型结合的方法,证明了这种方法更有利于半导体晶圆厂检测产品的各种特征参数。In this paper,a detection method of support vector machine(SVM)is proposed for feature parameter detection in semic on ductor manu facturi ng process.Firstly,the SVM model is an alyzed,and the results show that the two-dimensional SVM and the fuzzy SVM model are the best.Then a method combining the two models of 2D SVM and Fuzzy SVM is proposed,which proves that this method is more advantageous for semiconductor wafer factory to detect various characteristic parameters of products.

关 键 词：集成电路制造 机器学习 支持向量机SVM 特征参数检测 

分 类 号：TN405[电子电信—微电子学与固体电子学] TN407